Why does the concept of computing need a change before you revolutionize AI?

Computer and computing are related to machines that operate following detailed human instructions. However, AI refers to independent decision-making ability, and campaigns for that ability and we can never expect that feature from computers. People thought that by scaling up instructions for different situations we could create a machine that could respond to any situation that human encounters. A stupid would become intelligent by memorizing millions of facts and responding immediately. Even we started defining IQ in terms of how many problems a person could solve in a short span of time, speed and amount killed the quality of intelligence, everywhere. It means intellectual humanity thought that humans use limited reasoning, and limited sets of rules to live in this world and if a statistical database is large for the computer beyond a threshold number of instructions then a human would start believing that the machine responds like a human, so they are the same as far as intelligence is concerned.

The assumption here is that humans follow finite rules. What if humans follow a new kind of science where it synthesizes rules and the science of that synthesis is the road toward natural intelligence or a true AI? What if human consciousness holds the technology in creating non-repetitive decisions endlessly? What AI did the mistake is that they thought humans follow rules of reasoning like we write instructions for computers as algorithms. Thus far, faking a humanlike response by astronomically increasing instructions is already good enough and it is a matter of time, human-like responses would be there everywhere. However, the network of rules and integration of rules, synthesis of rules, are very different from an instruction follower. The blind instruction follower would fail in every step while showcasing the intricate details of execution.

Every moment human encounters an unprecedented situation and nature have made it fit to create something unprecedented. The rules of observed nature are a limit of sensory restriction of a human, and that limitation to make boundaries has made humans’ quest for knowledge limited to human observation. What that is not observed does not exist, is arrogance.

Just like “Moor’s law” or “singularity is near” predictions, here I make another prediction, one day, computers would have more instructions and rules for integrating information and yet they would make mistakes, they won’t have common sense and we would start concluding that algorithm like instructions are not the way human writes or processes instructions. My prediction is the existence of a crossover of limiting AI and biologically existing NI where it becomes clear that instructions do not have any chance further to zoom into human cognition and intelligence. This crossover is waited because of extraordinary promises of AI experts that astronomical trained responses, human-like mistakes, illogical expressions and emotion-filled contextual responses like humor, pain, anger and various human-only responses would come to an end when a real human finds it nonsense when the computer makes them. We cite three reasons below.

Our human subject experiments show that information has a very unique structure in the human mind and it’s not linear, it’s not non-linear. The event has a topological shape and they grow within and above. Then, we have been compiling instructions wrong all along. The engineering of phase explored by quantum mechanics is close but not suitable enough to simulate the brain. We need an engineering of singularity. We introduced geometric musical language, GML based on three fundamental elements, the engineering of singularity or undefined phase/amplitude, the engineering of infinity or undefined boundary, and the engineering of doping inside a singularity to grow information within and above. None of these features we have in quantum or classical computing. In fact, these brain features are orthogonal to current computing.

First, there is no classical state, neither states are defined nor is a finite value. The state of the mind is an endless chain of perceptions grown within and above. There is no reduction of choices or increment, it is a topological rearrangement of choices. Second, there is no black-and-white decision or choices. So, the brain’s options are clustered in a boundary, and the rapidly changing boundaries are the decision-making apparatus for the brain. Interchanging boundaries carry out the orthogonal transformation of a set of clocks localized in a boundary. The brain only finds invariants. Third, the brain seems to have a universal archive of invariants and integration rules. Possibly, this archive is fundamental to nature and allows living life forms to operate without requiring storing a huge amount of information. The brain does not require to acquire all facts and process information, rather it picks stored invariants and builds the invariant network, exclusive to the input information. Only those facts that make the terminal projections of the invariant network are absorbed in the brain and memorized.

Thus, philosophically brain is a device that unfortunately has no relation to the information theory. We need to find a way to make decisions without algorithms or instructions. A human mind-like core archive structure could be another way to build decision-making where an engine searches in the outside world for the invariants. Imagine what the difference would be. When our dead computers wait for inputs, the brain goes out and seeks more and more invariants by scanning the information. Then, the question arises that if we make a brain-like computer would it listen to us? The answer is yes. Just like quantum computer circuits, we may create points where the network could accept external input without disrupting the coherence. However, it would be a kind of computing where the machine hunts outside to evolve to be complete and at the same time, it tries to respond. Currently, machine learning claims to evolve by learning, but that’s packing more and more information and analyzing encoded protocols.

We agree that as time passes by AI would improve and get very close to humans but the three advantages of the brain that we have pointed out above need to be incorporated into the AI, because using large resources does not take us anywhere. The brain uses completely different technology to bypass reading entire information pixel by pixel. The brain does not follow the philosophy that the more the volume of information, the faster the speed of processing, then the better would be the performance. Shape, size, number of elements, speed, packing density all the parameters of computing become unimportant. The content resolution and content details are replaced by the context network. Information could be astronomical, symmetry is finite, and mapping breaking symmetry points reveals the architecture of relations between variables or recurring events.

Present-day computing means the reduction of choices and that’s why be it a classical logic gate or quantum, always, we have to find many choices and find reasons, to add, deduct, divide or multiply entities to reach a conclusion, that is simply rejecting many choices. The duality of choice and reason is a cyclic trap of computing, both increase together as the complexity of the problem increases. The more we increase the number of choices, the more we need to reject. Well, that is one way of making a decision. Also, earlier the choices were instant observations, recurrence and associations of recurring events and engineering of events, their true relationships were neglected totally. Moreover, logic was intuitions created by statistically dominating observations, along with a personal bias. So, artificial intelligence never wanted to learn how an event happens, but rather, how an event is observed. The idea was to increase the number of observations and make it a true expression of nature, a rule for those who have never observed it. Thus, the concept of present-day computer science has a serious flaw, it is architectures for the amount of information and then classifying it using an observational distinction. We think the way an event could happen is astronomical but it follows a limited number of fixed finite symmetries. Thus, converting information as a topology or geometric shape could represent astronomical variations.

It is something like a cube deformed in astronomical ways that could still be detected as a cube. Thus far no computing proposal was made where we could tell boldly, I have a limited set of geometric shapes and few rules to integrate them and any information could be rewritten in terms of a few geometric shapes bonded in a network. That geometry could withstand astronomical variations and errors in the input. Thus, training would not be important, it is rejecting everything fundamental to the existing information theory and replacing it with a fractal information theory, FIT.

Quantum computing is said to be probabilistic and with no defined state. In fact, it is not random. We divide infinite choices between zero and one but select a few choices with a phase theta and they are those decisions. We say but in reality, we do not explore astronomical choices in quantum computing. Popular videos make us fool they suggest those astronomical choices are harvested or used. That is never true. for n qubits, we take n phase angles and treat them as classical phase, i.e. defined states. So, eventually, its manipulation of phase that is profound in quantum computing. Quantum logic gate means selecting a few phase values of the states in the output. They are definitive, classical like in the mathematical presentation, we are often said that they are quantum only because, we assume that there is a link and a shared value of phase link among all possible values, much more profound than coupling. For classical coupling there is a factor, here for quantum we do not put a term called entanglement which has no parameter because it is not defined. The shared value of the tensor between two entangled states is not the measure of entanglement. Therefore we say with a clear conviction that the quantum algorithm is classical, all that is quantum about it is neither written nor could be written.

A Hadamard gate is a device where there is no classical output. If 1 is given as input, we get 0 and 1 superposition as output. When two Hadamard gates interact we get the classical state back. We could create many composite control gates by rotating the phase angle in the Bloch sphere, rotating the vector by an operation, and making the rotation of the vector follow various patterns, and the interaction of two Bloch spheres could lead to a classical state. Thus far quantum computing terminals were always a classical point. The concept of the mystery of quantum ends with those classical points. The assumption has been that classical points are reality.

Quantum logic gates are nothing but probability-modifying units. They don’t compute they simply alter phases of input states as instructed. Computing, therefore, has always been a method for the high-speed execution of a large volume of human instructions. Quantum is the next step of classical computing because the network or the entire circuit operates at once, and decoheres into the desired output. Quantum computing as it has been formulated do not hold any promise to make computer intelligent. The hunger for speed and resources has blinded the architect of quantum computers, even exploring the various possibilities of using hallmarks of quantum mechanics.

However, speed and resources are slaves of instructions, we need to learn to go beyond instructions, a universal treasure of keys of all keys that construct all information of the entire creation, from the smallest to the largest, extreme end of the cosmos. The magic box of codes of all codes should generate instructions in different layers following certain rules. The discovery of those rules would bring true artificial intelligence. A new generation of computing where we do not need any software engineers to use their intuition and write all possible instructions or create learning conditions. Thereafter, running millions of training from practical instances find critical bypass routes, and shortcuts, to link input and output. Faking intelligence with piles of black boxes would not be required anymore. Instructions would be pure science, an accurate mathematical representation of natural phenomena.

What if choices that are currently used in computations are replaced by combinations of mathematically defined all possible perceptions that could exist in the universe. We have long been working to find this key of keys. Prime is the fundamental element of the universe which attribute to no other elements. If we consider the prime number of choices to be the fundamentals of the information structure of the universe, the entire integer space embeds the choices of primes as architecture, this is the archive, the key of all keys. Therefore, we are replacing humans with a god-like archive that continuously creates keys for everything. Here everything means all key elements that could generate all possible facts in the universe. There could be another way of making decisions where we do not reduce the choices, we simply re-arrange them into an architecture that provides a much better and generic perspective about the situation.

Facts are projections of the continuously evolving invariant network of the universe. We are planning to change the century-old culture of acquiring knowledge. In brain-inspired computing we would not search for facts, rather, we would search for confusion, so that we could find more confusions inside a confusion. The journey through confusions would continue until we find only facts and fail to go inside. Phase singularity would carry the features of confusion. This new way of looking into the world was the symmetry of geometric shapes made by connecting the confusions to tell us which structure would form next or missing geometric shapes is a very abstract way to understand our universe. Generating geometric symmetry and creating infinite geometric shapes to create architectures do not include perceptions of nature, yet they deliver concepts that are very relevant to our nature. This unique journey is where blind connections of primes in the integers relate to geometric shapes and that relates to universal cognitive abilities and perceptions.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s